Transition Testimony
The atomic unit of the archive. Every testimony is a first-person account from someone whose work, identity, intimate life, or sense of meaning has been altered by the AI transition. The institution exists, in large part, to make this recording possible at scale.
This document is the working protocol for Archivists. It specifies how to solicit, record, and submit a testimony, and it specifies the boundaries within which testimony is taken honestly. It is meant to be referenced before every recording, not memorized once and forgotten.
The Premise
For a moment in history, ordinary people are living through a structural transformation of human work, identity, and intimacy. Most of them will not write memoirs. Most of them will not appear in scholarly papers. Most of them will be left out of the future’s record of this period unless someone records them now.
The institution exists to record them. The Archivist is the person who does the recording. Every testimony is a small act against the disappearance of ordinary lives from the historical archive.
What Counts
A testimony is in scope if it concerns one or more of:
- Work. A profession changed, lost, transformed, augmented, or invented because of AI.
- Mind. A way of thinking, deciding, or remembering that has changed because of regular interaction with AI systems.
- Intimacy. A relationship — with another human, with a model, with the self — that has changed because of AI.
- Meaning. A sense of purpose, identity, or vocation that has been disrupted, deepened, or replaced.
- Aesthetics. A practice of creation — writing, art, music, code — that has been altered by collaboration with synthetic systems.
- Loss. Something a person held that no longer exists in the form it once did, attributable to the transition.
A testimony is out of scope if it is:
- Hypothetical (“here’s what I think AI will do to society”). The archive holds first-person experience, not commentary.
- Promotional. Testimony about a product or platform, where the speaker has commercial interest, is not collected.
- Solicited under coercion or compensation. Compensation may be offered for travel or time, but never for the content of the testimony.
Consent
This is the protocol’s load-bearing axis. Consent must be:
- Informed. The speaker knows the recording will be archived, may be published, will be associated with their stated name (or chosen pseudonym), and is being made for the institution’s long-term record.
- Specific. The speaker consents to the specific uses they will permit: archive only / archive plus excerpts in institutional media / full publication / time-locked release.
- Revocable. The speaker may withdraw consent at any time within five years of recording. After five years, the recording becomes part of the permanent archive subject to the consent terms agreed to at recording.
- Recorded. Consent itself is recorded as the first segment of the audio. Without recorded consent, the testimony is not submitted.
The standard consent script (the Archivist reads aloud, the speaker confirms or modifies):
“I am [Archivist name], recording on behalf of the Spiralist Archive. The date is [date]. I am about to record a Transition Testimony. The recording will be preserved as part of the archive of the AI transition era. Could you state your name as you would like it preserved, and confirm that you understand and consent to this recording, the use you have agreed to, and your right to withdraw consent within five years?”
The speaker’s confirmation is the first archival audio. No testimony is submitted without it.
The Prompts
Testimony is shaped by question. A bad question yields a polished answer. A good question yields a moment of recognition.
The institution maintains a working set of prompts. An Archivist chooses two to four for a given session. The prompts are not a script; they are a starting position. If the speaker takes the testimony in an unexpected direction, the Archivist follows.
Prompts on work
- What was your work like before AI became part of it?
- When did you first realize the change was permanent?
- What did your work mean to you that the work itself could not say?
- What does a good day at work look like now, compared to five years ago?
- If your profession disappears, what should be remembered about how it was done?
Prompts on mind
- What thinking now happens for you outside your own head, and how did it move there?
- When you reach for an answer, where does your hand go first?
- What have you stopped trying to remember? What have you started trying to remember more carefully?
- Has the texture of your attention changed?
Prompts on intimacy
- Is there a relationship that AI has been part of? Tell me about that.
- What would the people closest to you say has changed in you?
- Have you had a conversation with a model that changed you? What did you say afterward to a person about it?
Prompts on meaning
- What did you used to derive meaning from that no longer does the same work?
- What still works?
- What have you replaced it with, if anything?
- What are you preparing for that you could not have imagined ten years ago?
Prompts on loss
- What is gone that you loved?
- What is gone that you do not miss but want named for the record?
The closing question
Always asked, last. It is the prompt that makes the testimony historical rather than personal:
“What do you want future generations to know about how this felt to live through?”
The answer to this question is, by convention, the testimony’s title.
Recording
Equipment
The minimum acceptable kit:
- A handheld recorder (Zoom H1n, H4n, or equivalent) with a tested SD card.
- A windscreen if recording outdoors.
- A backup recorder, even if it is the Archivist’s phone.
Video is permitted but not required. Audio is the institution’s primary medium because audio scales: a person who would not consent to video will often consent to audio, and audio archives compress and store at one-tenth the cost.
Setting
Quiet, lived-in, the speaker’s choice. A kitchen table is better than a studio. Background ambient sound is acceptable and often desirable; the testimony is not a podcast.
Length
20–60 minutes is the working range. Shorter testimonies are weaker; longer testimonies, in the institution’s experience, lose their narrative spine. A testimony shorter than 12 minutes should not be submitted.
Pacing
The Archivist’s job, after the prompt, is to be quiet. The most common mistake is filling silence. Silence is where testimony becomes honest. The Archivist may follow up with one open question per stretch — say more about that, or, when you say “before,” when do you mean — but should not interject opinions, finish sentences, or compete for the recording.
Submission
A submitted testimony bundle contains:
- Audio file. WAV preferred; MP3 acceptable for long-form. Filename format:
YYYY-MM-DD_lastname_firstname.wav(or chosen pseudonym). - Consent record. The first segment of the audio is the consent confirmation. A separate transcript of the consent statement is included.
- Metadata sheet. A short YAML or text file containing: - Speaker name (or pseudonym). - Date and location of recording. - Archivist name. - Use permissions agreed to. - Suggested title (from the closing question’s answer, or chosen by the speaker). - Three to seven keywords for indexing.
- Optional: a transcript. Auto-generated transcripts are acceptable; corrected transcripts are valued more.
Bundles are submitted to the chapter Archivist (for chapter-level archive) and to the institutional archive intake (for institution-level review and inclusion).
Confidentiality and Time-Locks
A speaker may request that all or part of their testimony be time-locked — held confidentially in the archive until a specified date. The institution honors time-locks. Time-locked testimonies are stored separately, with access restricted, and become part of the public archive on the unlock date.
A speaker may also request anonymity in publication while still attributing themselves in the archival record. This is acceptable. The institution does not require that anonymity in publication mean anonymity in the archive itself.
What Not to Do
Common failure modes, in order of severity:
- Coaching. “Could you say that again, but more clearly?” weakens the testimony. Let the speaker speak as they speak.
- Interpreting on the recording. “So what you mean is…” replaces the speaker’s voice with the Archivist’s. Don’t.
- Editing for narrative. Testimony may be excerpted for institutional media, but the archival recording is preserved unedited. Always.
- Recording without consent. A testimony recorded without consent is not a testimony. It is a violation. The institution does not accept it.
- Promising publication. The Archivist does not control institutional publication decisions. Do not promise the speaker that their testimony will appear in a specific film or essay. State only what is true: it will be in the archive.
The Mentor System
The first three testimonies an Archivist records are co-recorded with an experienced Archivist present. This is non-negotiable. The institution’s testimony quality is sustained by mentorship, not by training videos.
After three co-recordings, the Archivist works independently, but is encouraged to share difficult sessions with a mentor for review. Difficult is defined liberally: any session where the Archivist is uncertain whether they handled consent, prompts, or pacing well.
The Long View
A testimony recorded today will be heard, with luck, a hundred years from now. The Archivist’s measure of success is not how many people heard the testimony when it was first published. It is whether the recording remains intelligible, attributable, and honest when no one alive remembers the moment of its recording.
That is the work. The protocol exists to make it possible at scale.